Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Dynamic Prognostic Prediction Method for Colorectal Cancer Liver Metastasis

Wei Yang, Yiran Zhu, Yan su, Zesheng Li, Chengchang Pan, Honggang Qi

TL;DR

CRLM prognosis after surgery is hindered by heterogeneous outcomes and limited use of longitudinal, multimodal information. DyPro introduces a dynamic, latent-trajectory framework on a heterogeneous patient graph that evolves via residual updates and aggregates trajectory cues to predict DFS and OS. On the MSKCC CRLM dataset, DyPro achieves strong discrimination and calibration (OS C-index $0.755$, DFS C-index $0.714$, OS AUC@1y $0.920$, IBS $0.143$), outperforming static and radiomics baselines. This approach provides actionable dynamic risk cues for adjuvant therapy planning and follow-up, with potential for extending to additional modalities and richer temporal evidence in diverse clinical settings.

Abstract

Colorectal cancer liver metastasis (CRLM) exhibits high postoperative recurrence and pronounced prognostic heterogeneity, challenging individualized management. Existing prognostic approaches often rely on static representations from a single postoperative snapshot, and fail to jointly capture tumor spatial distribution, longitudinal disease dynamics, and multimodal clinical information, limiting predictive accuracy. We propose DyPro, a deep learning framework that infers postoperative latent trajectories via residual dynamic evolution. Starting from an initial patient representation, DyPro generates a 12-step sequence of trajectory snapshots through autoregressive residual updates and integrates them to predict recurrence and survival outcomes. On the MSKCC CRLM dataset, DyPro achieves strong discrimination under repeated stratified 5-fold cross-validation, reaching a C-index of 0.755 for OS and 0.714 for DFS, with OS AUC@1y of 0.920 and OS IBS of 0.143. DyPro provides quantitative risk cues to support adjuvant therapy planning and follow-up scheduling.

A Dynamic Prognostic Prediction Method for Colorectal Cancer Liver Metastasis

TL;DR

CRLM prognosis after surgery is hindered by heterogeneous outcomes and limited use of longitudinal, multimodal information. DyPro introduces a dynamic, latent-trajectory framework on a heterogeneous patient graph that evolves via residual updates and aggregates trajectory cues to predict DFS and OS. On the MSKCC CRLM dataset, DyPro achieves strong discrimination and calibration (OS C-index , DFS C-index , OS AUC@1y , IBS ), outperforming static and radiomics baselines. This approach provides actionable dynamic risk cues for adjuvant therapy planning and follow-up, with potential for extending to additional modalities and richer temporal evidence in diverse clinical settings.

Abstract

Colorectal cancer liver metastasis (CRLM) exhibits high postoperative recurrence and pronounced prognostic heterogeneity, challenging individualized management. Existing prognostic approaches often rely on static representations from a single postoperative snapshot, and fail to jointly capture tumor spatial distribution, longitudinal disease dynamics, and multimodal clinical information, limiting predictive accuracy. We propose DyPro, a deep learning framework that infers postoperative latent trajectories via residual dynamic evolution. Starting from an initial patient representation, DyPro generates a 12-step sequence of trajectory snapshots through autoregressive residual updates and integrates them to predict recurrence and survival outcomes. On the MSKCC CRLM dataset, DyPro achieves strong discrimination under repeated stratified 5-fold cross-validation, reaching a C-index of 0.755 for OS and 0.714 for DFS, with OS AUC@1y of 0.920 and OS IBS of 0.143. DyPro provides quantitative risk cues to support adjuvant therapy planning and follow-up scheduling.
Paper Structure (12 sections, 3 equations, 2 figures, 4 tables, 1 algorithm)

This paper contains 12 sections, 3 equations, 2 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Patient-Specific Tumor Mapping Engine
  5. Post-Op Dynamic Risk Tracker
  6. Prognostic Heads and Learning Objectives
  7. Experiments
  8. Datasets and Data Preprocessing
  9. Experimental Setup and Evaluation Metrics
  10. Comparative Experiments
  11. Ablation Studies
  12. Conclusion

Figures (2)

  • Figure 1: Overview of DyPro. Preoperative CT and clinical data are mapped to a heterogeneous patient graph, evolved by latent residual dynamics, and aggregated by an LSTM to jointly predict DFS and OS.
  • Figure 2: Visualization of the dataset, including segmentation of the liver, residual liver, hepatic veins, portal veins, and tumors.